Using AI Intentionally: A Systems Thinking Approach

For the past three years, people have asked which AI tools to use, what prompts work best, and how to 'get ahead' with it. But the real opportunity isn't in tools or tricks. It's in learning how to think differently - how to integrate AI into everyday work, learning, and creativity. This essay offers a mental model for doing just that - so you can use AI more intentionally and effectively.

One of the biggest barriers to using AI well is how we've been taught to think. Most of us grow up learning how to succeed in structured, rule-based systems - school, exams, standardized goals. We're trained to look for shortcuts, formulas, and ways to win. Paul Graham calls this "playing house" - performing the motions without building anything real.

People approach AI the same way. They want the cheat code. But there isn't one. AI isn't a game. It's a system of tools, behaviors, and interactions that shape how you think and what you create. This essay offers a different approach: not tactics, but a mental model.

We'll get there by exploring bathtubs (yes, really), mental models, and the role of intentional friction.

A Quick Primer on Systems Thinking

"In our world," said Eustace, "a star is a huge ball of flaming gas." "Even in your world, my son, that is not what a star is, but only what it is made of."

• C.S. Lewis, The Voyage of the Dawn Treader

To frame how AI fits into our lives, I draw from Thinking in Systems by Donella Meadows - a book that explains how complex things work, whether it's an economy, a climate system, or how we interact with AI.

Systems thinking asks: how do things accumulate, reinforce, and balance over time? It shifts us from cause-effect logic ("X leads to Y") to patterns, loops, delays, and - most relevant here - stocks and flows.

Meadows uses a simple metaphor: a bathtub. The water already in it is the stock. The faucet is the inflow. The drain is the outflow. The level depends on how fast water enters and exits.

This idea applies beyond plumbing. It explains how energy builds, how habits form, how momentum stalls. Stocks shift slowly, even if flows are rapid. And while flows are easy to miss, they quietly shape everything.

So what exactly is a stock? It's anything that builds up over time. In your personal system, that might mean your knowledge, emotional resilience, creative confidence, or decision-making clarity. These aren't instantly generated - they're slowly cultivated through repeated exposure, reflection, and reinforcement.

That's the lens we'll use. Because when it comes to AI, what we take in, retain, and express - our inflows, stocks, and outflows - define how deeply and intentionally we use it. In fact, I'm going to use it as a lens for thinking about AI. Because when you zoom out, the way we use AI is deeply tied to our own inflows, stocks, and outflows - what we take in, what we store or internalize, and what we express or act on.

Inflows: AI as a Curator

What we take in shapes how we think - and most of what we take in is shaped for us. Newsfeeds, homepages, social media: they're curated to capture attention, not deepen understanding. As Elon Musk even admits this in a post, "the [X/Twitter] algorithm is trying to maximize probable long-term user-seconds." Ultimately, these systems care more about what will keep you engaged - not what's true or useful.

AI changes that. It gives you the ability to actively seek information based on your interests - not what's trending, not what a platform wants you to click. You can dig into new recipes, get a plain-English breakdown of reinforcement learning, explore niche topics, or ask for tailored advice. AI rewards curiosity.

This is a rare opportunity to take control of your inflows. To choose what enters your mind, instead of letting the world decide for you.

Remember the bathtub: you control the faucet. What you choose to let in determines what accumulates over time.

Stocks: AI as a Thinking Partner

If inflows are what we take in, stocks are what we hold onto.

In systems thinking, a stock is anything that accumulates over time - knowledge, beliefs, mental models, creative energy. These are the reservoirs we draw from when we act, write, decide, or create.

What's tricky is that stocks change slowly. You can read five great essays or watch a dozen videos in a day, but that doesn't mean the ideas stick or that they reshape how you think. Real accumulation takes time, reflection, and engagement.

AI can serve as a kind of second brain - an extension of internal thought. Like an alternate inner monologue, it reflects, refines, and helps shape ideas as they take form. Use it to externalize messy, half-formed ideas - capturing voice-to-text thoughts, raw reactions, or rough outlines - and then refine them with the help of the model. It's like having a responsive mirror for thought.

This process of externalizing, rephrasing, and pattern-finding is particularly useful when working through something abstract or complex. It can help clarify what's already understood, surface what isn't, and bring form to something still taking shape. In that sense, it accelerates the formation of stock - deep, internal knowledge - rather than simply accelerating output.

But there's a risk. AI makes it easy to skip the slow, reflective part. Without intention, the process can become surface-level. Instead of building understanding, it just polishes fragments.

Used well, AI supports reflection. It helps slow down thinking, not just speed it up. And that's where the real growth happens.

Outflows: AI as an Amplifier

If inflows are what we take in and stocks are what we retain, then outflows are what we express - our writing, decisions, designs, conversations, actions, or creative output. It's how internal understanding turns into external impact.

This is where AI really shines for many people: it makes doing easier.

Whether you're writing an email, drafting a proposal, brainstorming ideas, or prototyping code, AI tools can dramatically reduce friction between thought and action. You don't have to start from scratch. You can get a first draft in seconds. You can translate vague ideas into polished text, or take a dry outline and make it engaging.

But as with inflows and stocks, there's a risk in going passive: letting AI become the voice instead of the amplifier. If you're not careful, you can end up outsourcing not just the writing, but the thinking behind it. The result might sound polished, but it lacks soul. You're not drawing from your stock - you're bypassing it.

Used well, AI lets you scale your voice without losing it. It lets you express more of what's in you. But only if what's in you - the stock - is real.

Pitfalls, Friction, and the Hardest Use Case

Powerful systems run quickly - but that speed can be dangerous.

The world around us is increasingly designed to remove friction. Infinite scroll. One-click ordering. Default recommendations. AI fits right into this trend. It makes things smoother, easier, faster. But without friction, you're not really choosing - you're just consuming.

This is why I think adding intentional friction is one of the best things you can do. Ask AI to explain things multiple ways. Try rewording your own ideas before asking for help. Take notes. Pause to reflect.

And recognize that not all AI use is equal.

Most people engage with AI at three levels:

Outflows – Using AI to write, code, summarize, polish. This is the most intuitive and widely adopted.

Inflows – Using AI to learn, research, explore. Slightly deeper, but still common.

Stocks – Using AI to think better, to wrestle with ideas, to build real internal understanding. This is rare. It takes more effort - but it's the most valuable.

If you want to grow your creative, intellectual, or emotional reservoir - your stock - then use AI to support that goal. Don't just read faster or write more. Think deeper and be curious.

Closing: Use the System, Don't Be Used by It

AI isn't just a tool - it's part of the system you live in. It shapes what you see, what you think, and what you put into the world. The more aware you are of that system - your inflows, your stock, your outflows - the more intentional and powerful your use of AI becomes.

The goal isn't just to move faster. It's to move smarter, with more clarity, more direction, and more ownership. Remember, the most powerful tools don't change what we do. They change what we believe is possible.